JPWO2019216235A1 - Angle detector - Google Patents

Abstract

The present invention provides an angle detector that detects an amount of angle change due to rotation of a rotating body. The angle detector 101 includes a rotating body 105 that rotates about a rotation axis, a scale scale 102 having a plurality of scales 103 along the circumference of the rotating body in the rotation direction, and a plurality of sensors arranged along the circumference. Each sensor 201a and 201b includes 201a and 201b, and each sensor 201a and 201b outputs a signal according to the amount of angle change based on the plurality of scales 103, and the output signal is one scale of the plurality of scales 103 in one cycle 1 The amount of angular displacement calculated from the output signal, which includes the following fundamental wave component and the harmonic component whose order is an integral multiple of 2 or more of the fundamental wave component, is one scale due to the harmonic component. The number of the plurality of sensors 201a and 201b is determined based on the scale number of the scale scale and the order component of the angle error component.

Description

The present invention relates to an angle detector for detecting an amount of change in an angle due to a rotational motion of a rotating body.

Angle detectors such as encoders, resolvers, and induct thins are used to detect the amount of angle change due to the rotational movement of the rotating body. The angle detector includes a scale scale on which a plurality of scales are arranged, a sensor that reads the plurality of scales, and a control unit that converts information read from the sensor into an amount of change in the angle of a rotating body. Either the scale scale or the sensor is attached to the rotating body. In order to read the amount of change in the angle of the rotating body with higher resolution, the angle interval of one scale on the scale scale may be narrowed, but the scale cannot be made infinitely fine because it is carved by processing, for example. In order to measure the amount of change in the angle of the rotating body in more detail, there is a method in which the output signal based on the information read from the sensor is numerically calculated by the control unit and one scale is finely divided. The output signal from the sensor used in the angle detector is usually in the form of a square wave or a sine wave, and is often a two-phase signal having one scale of 360 ° per period and a phase difference of 90 °. .. When the output signal from the sensor is a sine wave signal, the two-phase signal has the shape of cosθ and sinθ with one scale as one cycle. As a method of dividing one scale, for example, a method of performing an inverse tangent operation on a two-phase signal can be mentioned (that is, θ = tan ^-1 (sin θ / cos θ)). In this method, the angular resolution can be improved according to the detection resolution of the amplitude of the output signal from the sensor. However, since the output signal from the sensor contains distortion of harmonic components in addition to the ideal one-scale, one-cycle sine wave signal, not only the true amount of change in the angle of the rotating body, but also Depending on the distortion of the harmonic component, an amount irrelevant to the true amount of angle change is also included in the measurement. That is, between the true angle change amount due to the rotation of the rotating body based on the command and the angle change amount due to the measurement obtained by converting the output signal from the sensor by the control unit, the true amount due to the influence of the harmonic component An error (angle error) that is irrelevant to the amount of angle change occurs. In order to measure the true amount of angle change and angle error, it is necessary to remove the distortion of the harmonic component contained in the output signal from the sensor.

Further, in the angle detector, it is desirable that the rotation axis of the rotating body and the center axis of the scale scale are arranged on the coaxial line. However, in general, these axes do not exactly match, and the offset between these axes (axis eccentricity) causes the true angle change due to the rotation of the rotating body based on the command and the output from the sensor. An angle error occurs between the signal and the amount of change in angle due to measurement obtained by converting the signal by the control unit. Further, since the scale of the scale scale is engraved by processing, for example, the center of the scale pattern and the rotation center of the scale scale itself are offset, and the interval of one scale has an error with respect to the ideal value, which is not possible in a plurality of scales. There are quality problems such as equality. In addition, the detection accuracy of the angle detector itself changes over time as the components deteriorate. Therefore, the control unit also converts the true amount of angle change due to the rotation of the rotating body and the output signal from the sensor. There is an angle error with the amount of change in angle due to the measurement obtained.

Due to such a problem, Patent Document 1 discloses a method of detecting and removing distortion of a third-order harmonic component included in a two-phase sinusoidal signal having a phase difference of 90 °, and Patent Document 2 discloses. A method for detecting and removing distortions of harmonic components of the third and fifth orders contained in a two-phase sinusoidal signal having a phase difference of 90 ° is disclosed.

Further, Patent Documents 3 to 5 include a plurality of first scale reading heads and one second scale reading head around a scale board fixed to the rotation axis, and the second scale reading head is provided as described below. angle signals a _{i, 1} and the angle signal a _i of the first graduation readhead scale read _head, the difference between the _j SA _i, the angle detector for performing self-calibration by obtaining an average value SAV _i seeking _j is It is disclosed. Here, i is a scale number (integer of _{1 to NG , NG} is the total number of scales), j is a scale reading head number (integer of _{1 to NH , NH} is the total number of scale reading heads).

Japanese Unexamined Patent Publication No. 2006-112862 Japanese Unexamined Patent Publication No. 2008-304249 Japanese Unexamined Patent Publication No. 2006-98392 Japanese Unexamined Patent Publication No. 2011-99002 Japanese Unexamined Patent Publication No. 2011-99004

In Patent Documents 1 and 2, the electrical angle error due to the distortion of a specific harmonic component caused by the sensor can be detected and removed. However, the output signal from the sensor contains distortion of various harmonic components depending on the accuracy of the scale, the characteristics and method of the sensor, etc., and it is not possible to detect the distortion of a specific harmonic component. It is not possible to remove the distortion of the harmonic component uniformly for the sensor. For example, the strain characteristics are completely different between the optical and magnetic reading methods of the sensor, and the strain characteristics are completely different between the magnetized ring and the gear, which are read by the magnetic type. Further, when a device such as an amplifier that amplifies the output signal from the sensor is used, distortion of harmonic components also occurs due to those characteristics. Therefore, in some cases, the distortion of the harmonic component can be removed, but in another case, the distortion of the harmonic component cannot be removed.

In Patent Documents 3 to 5, the angle detector detects the mechanical angle error caused by the axis eccentricity between the rotation axis of the rotating body and the central axis of the scale scale, the quality of the scale scale, the aging of the angle detector, and the like. Can be removed. In order to obtain the amount of change in the angle of the rotating body with higher accuracy using this method, it is effective to narrow the angle interval of one scale or to use a highly accurate scale reading head. It tends to lead to higher cost of the detector. However, if a scale scale with wide scale intervals or a scale reading head with low accuracy is used to reduce the cost, the amount of harmonic components contained in the signal output by the scale reading head will increase the amount of influence on the electrical angle error. Since it appears as a dominant error rather than the mechanical angle error, it is not possible to obtain the amount of change in the angle of the rotating body with high accuracy.

Therefore, an object of the present invention is to simultaneously remove not only the electrical angle error caused by the sensor but also the mechanical angle error caused by the mounting accuracy of the rotating body, the quality of the scale scale, the change over time of the angle detector, and the like. Therefore, it is an object of the present invention to provide an angle detector capable of obtaining the amount of change in the angle of the rotating body with high accuracy and reducing the cost.

According to one aspect of the present invention, the angle detector includes a rotating body that rotates about the axis of rotation, a scale scale that has a plurality of scales along the circumference of the rotating body with respect to the rotation direction, and a rotation direction of the rotating body. It is equipped with at least two sensors arranged along the circumference of the rotating body to detect the amount of change in angle due to the rotation of the rotating body, and each of the at least two sensors is a signal according to the amount of change in angle based on a plurality of scales. Is output, and the output signal includes a fundamental wave component having one scale of a plurality of scales as the first order in one cycle and a harmonic component having an integral multiple of two or more of the fundamental wave components as the order, and is output. The amount of angular displacement calculated from the signal includes at least one angular error component having an order component that is an integral multiple of one scale with one cycle as the first order due to the harmonic component, and of at least two sensors. The number is determined based on the number of scales on the scale scale and the order of at least one angular error component.

According to a specific example of the present invention, in the angle detector, at least one angle error component is a plurality of angle error components, and the number of at least two sensors is the number of scales on the scale scale and the plurality of angle errors. It is determined based on the order of each of the components.

According to a specific example of the present invention, in an angle detector, the number of at least two sensors is divisible by the product of the number of scales on the scale scale and an integer of one or more, which is the order of at least one angle error component. It is determined based on an integer that cannot be used.

According to a specific example of the present invention, in an angle detector, the number of at least two sensors is too much when divided by an integer that cannot divide the product of the scale number of the scale scale and the order of the angle error component. It is decided based on.

According to a specific example of the present invention, in the angle detector, the number of at least two sensors is the remainder when the product of the scale number of the scale scale and the order of the angle error component is further divided by an integer that cannot be divided. It is determined based on the weighting according to the size of.

According to a specific example of the present invention, in the angle detector, a number of planned sensor placement locations corresponding to an indivisible integer are set at approximately equal intervals around the rotation direction of the rotating body, and the sensor placement planned locations are set. At least two sensors are arranged one by one in any of the above.

According to a specific example of the present invention, in the angle detector, at least two sensors are arranged at two adjacent positions where the sensors are planned to be arranged.

According to a specific example of the present invention, in the angle detector, at least two sensors are arranged at each of the planned sensor arrangement locations.

According to a specific example of the present invention, in an angle detector, at least one angle error component differs depending on the type of at least two sensors.

According to a specific example of the present invention, the angle detector self by finding the output signal difference between the output signal from at least one of the two sensors and the output signal from the other sensor. It has been calibrated.

According to the present invention, the angle detector is caused by an electrical angle error due to distortion contained in an output signal from the sensor, mounting accuracy of a rotating body, quality of scale scale, aging of the angle detector, and the like. The amount of change in the angle of the rotating body can be obtained with high accuracy by simultaneously removing the mechanical angle error, and the angle detector can change the angle of the rotating body with high accuracy without being caused by the sensor. The amount can be obtained. Further, according to the present invention, the accuracy of the angle change amount of the rotating body can be significantly improved by using an inexpensive and low-precision sensor, and it is not necessary to use a conventional expensive and highly accurate sensor, and the angle detector is used. Cost reduction can be achieved.

It should be noted that other objects, features and advantages of the present invention will become clear from the following description of the examples of the present invention with respect to the accompanying drawings.

It is a schematic diagram which shows the angle detector which detects the angle change amount of the rotating body which attached the scale scale as one Embodiment of this invention. It is the schematic which shows the angle detector which detects the angle change amount of the rotating body which attached the scale scale as another embodiment of this invention. It is the schematic which shows the angle detector which detects the angle change amount of the rotating body which attached at least two sensors as another embodiment of this invention. It is the schematic which shows the angle detector which detects the angle change amount of the rotating body which attached at least two sensors as another embodiment of this invention. It is the schematic of the control part of the angle detector of FIGS. 1A-2B. The pseudo sine wave signal output when the sensor detects the position in one scale is shown. The measurement angle and the ideal angle with respect to the position in one scale calculated by the pseudo sine wave signal of FIG. 4A are shown. The angle error between the measurement angle and the ideal angle with respect to the position in one scale calculated by the pseudo sine wave signal of FIG. 4A is shown. It is a table which shows the result of having made the scale number constant, dividing the angle error order around the scale scale by the number of places where a sensor is planned, and making a judgment based on the fact that it is divisible / not divisible. It is a table which shows the result of having determined based on the weighting according to the size of the remainder when the angle error order around the scale scale is divided by the number of places where a sensor is planned to be arranged with the number of scales is constant. It is a table which shows the result of having determined based on the weighting according to the size of the remainder when the angle error order around the scale scale is divided by the number of places where a sensor is planned to be arranged with the number of places where a sensor is planned to be arranged constant. It shows the angle error with respect to the command angle of the rotating body before the electrical angle error by the sensor and the mechanical angle error due to the mounting accuracy of the rotating body are removed. It is an enlarged view of a part of FIG. 6A when the command angle of a rotating body is converted into a scale. It was obtained by executing the Fourier transform on the angle error in which the circumference of the rotating body was set to the first order in one cycle before the electrical angle error by the sensor and the mechanical angle error due to the mounting accuracy of the rotating body were removed. Shows the spectral intensity. The angle error with respect to the command angle of the rotating body after the electrical angle error by the sensor and the mechanical angle error due to the mounting accuracy of the rotating body are removed is shown. It is an enlarged view of a part of FIG. 7A when the command angle of a rotating body is converted into a scale. It was obtained by executing the Fourier transform on the angle error in which the circumference of the rotating body is set to the first order in one cycle after the electrical angle error by the sensor and the mechanical angle error due to the mounting accuracy of the rotating body are removed. Shows the spectral intensity.

Hereinafter, examples of the present invention will be described with reference to the drawings, but the present invention is not limited to these examples.

1A to 2B show a rotating body 105 rotating about the rotation axis 108, a scale scale 102 having a plurality of scales 103 along the circumference of the rotating body 105 with respect to the rotation direction 106, and a rotation direction 106 of the rotating body 105. Indicates an angle detector 101 including at least two sensors 201a to 201i arranged at planned sensor placement locations 202a to 202i. Here, the scale 103 included in the scale scale 102 is not only one that can be visually recognized as if it was actually engraved on the scale scale 102 by processing, for example, but also a predetermined position interval on the scale scale 102 for one scale. Anything that can be read by the sensors 201a to 201i as an interval is sufficient, and the scale scale 102 is a member in which a plurality of such scales 103 are arranged. Typical examples of the angle detector 101 include an encoder, a resolver, an induct thin, and the like, but the principle is not particularly limited as long as the present invention can be applied. Further, the principle of the sensors 201a to 201i is not particularly limited as long as the scale 103 of the scale scale 102 can be read. Examples of the sensors 201a to 201i include an optical sensor, a magnetic sensor, a coil, and the like. Further, the scale scale 102 may be any as long as the sensors 201a to 201i can read the scale 103, and the material, the method of arranging the scale 103, and the like are not limited. The sensors 201a to 201i are arranged around the circumference 107 of the rotating body 105 with respect to the rotation direction 106, that is, at the planned sensor placement locations 202a to 202i along the entire circumference. Angle detector 101, the angle variation _{X p} by the rotation of the rotating body 105, the detection is based on the output signal from the sensor 201a~201i by using a plurality of graduations 103 which are arranged along the rotational direction 106 .. Each of the sensors 201a to 201i outputs a signal 204 according to the amount of change in the angle due to the rotation of the rotating body 105 based on the plurality of scales 103. Here, the width between two adjacent scales 103, that is, the angle interval 104 of one scale is indicated by X.

FIG 1A, 2 two sensors 201a along the circumference 107 of the rotary member 105, 201b is arranged, detects an angle change amount _{X p} of the rotating body 105 to rotate in the rotational direction 106 of graduated scale 102 is attached The angle detector 101 is shown, and in FIG. 1B, nine sensors 201a to 201i are arranged along the circumference 107 of the rotating body 105, and the rotation is rotated along the rotation direction 106 to which the scale scale 102 is attached. indicated angle detector 101 for detecting an angle change amount _{X p} body 105, FIG. 2A, along the circumference 107 of the rotary member 105, three sensors 201a, 201b, 201d are arranged, three sensors 201a , 201b, the angle detector 101 for detecting an angle change amount _{X p} of the rotating body 105 to rotate in the rotational direction 106 201d is attached is shown in the Figure 2B, along the circumference 107 of the rotary member 105 are arranged nine sensors 201A～201i, the angle detector 101 for detecting an angle change amount _{X p} of the rotating body 105 to rotate in the rotational direction 106 of the nine sensors 201A～201i is attached is shown.

As shown in FIG. 3, the angle detector 101 further connected to the sensor 201A～201i, a control unit 203 for converting the read information of the sensor 201A～201i the angle variation _{X p} of the rotary body 105. The converted angle change amount _Xp may be output to the display device 211, etc., or may be fed back to the motor for driving the rotating body 105, the control device for the rotating body 105, and the like.

ここで、θ^（０）＝ｔａｎ^−１（Ｂ^（０）／Ａ^（０））は０〜２πの範囲となるように数値処理されている。また目盛検出数Ｍは、θ^（０）＝ｔａｎ^−１（Ｂ^（０）／Ａ^（０））が０と２πの境界を超えるタイミングで、そのカウント値を増減させる等の処理で検出可能であって、その方法については問われない。図１Ａ〜図２Ｂのように、目盛スケール１０２又はセンサ２０１ａ〜２０１ｉを回転体１０５に取り付けて回転させる場合には、制御部２０３は、回転体１０５の回転による角度変位量Ｘ_ｐとして算出できる。Sensors 201a to 201i generally cycle one of a plurality of scales 103 based on the read scale 103 and the angular distance 104 of one scale when a relative movement with the scale scale 102 occurs. It is possible to output an output signal 204 which is primary and whose amplitude changes according to the amount of change in angle. The control unit 203 includes an output signal 204 from the sensor 201A～201i, and a scale detection number M counted until a certain time, it can be converted into angle variation _{X p} of the rotary body 105. As shown in FIGS. 1A to 2B, due to the relative motion between the scale scale 102 and the sensors 201a to 201i, for example, two pseudo sine waves having 90 ° different phases as shown in FIG. 5A from the sensors 201a to 201i, respectively. When the signal (A-phase signal (A ⁽⁰⁾ ) and B-phase signal (B ⁽⁰⁾ )) is output as the output signal 204, the control unit 203 uses the pseudo sine wave output by each of the sensors 201a to 201i. The signal is acquired, and one of the two pseudo-sine wave signals from the sensors 201a to 201i whose phase is delayed by 90 ° (B-phase signal (B ⁽⁰⁾ )) is replaced with the other pseudo-sine wave signal. (A phase signal (A ⁽⁰⁾ ) divided by the inverse tangent calculation is performed to divide one scale, and the provisional angle change amount X _p ⁽⁰⁾ of the rotating body 105 is calculated as follows.

Here, θ ⁽⁰⁾ = tan ^-1 (B ⁽⁰⁾ / A ⁽⁰⁾ ) is numerically processed so as to be in the range of 0 to 2π. The scale detection number M can be detected by processing such as increasing or decreasing the count value at the timing ^{when θ (0)} = tan ^-1 (B ⁽⁰⁾ / A ^{(0)) exceeds the boundary between 0 and 2π.} There is no question about how to do it. As shown in FIG 1A~ Figure 2B, the case of rotating by attaching the graduated scale 102 or sensor 201a~201i the rotating body 105, the control unit 203 can be calculated as the angle displacement amount _{X p} by the rotation of the rotary member 105.

ここで、ａ_ｋ及びｂ_ｋは、１目盛を１周期とする１次の基本波成分の振幅を１とした場合の次数ｋの高調波成分のゲインであって、φａ_ｋ及びφｂ_ｋは、次数ｋの高調波成分の基本波成分に対する位相差である。なお、ａ_ｋ、ｂ_ｋ、φａ_ｋ、φｂ_ｋは、一般的には、異なる目盛１０３であっても変化しないか、又は変化したとしても小さい相違である。なお、ａ_ｋ、ｂ_ｋ、φａ_ｋ、φｂ_ｋは、センサ２０１ａ〜２０１ｉや目盛スケール１０２の特性、検出原理により決定付けられ、例えば、センサ２０１ａ〜２０１ｉが光検出型センサであれば目盛スケール１０２の目盛パターンの反射・透過特性や受光部の感度特性により決定付けられ、センサ２０１ａ〜２０１ｉが半導体磁気抵抗型センサであれば半導体の磁気抵抗特性により決定付けられ、また、半導体磁気抵抗型センサを使用する場合において、磁気検出に平歯車を使用すると、歯の形状特性により決定付けられる。このように、出力信号２０４に含まれる高調波成分は、センサ２０１ａ〜２０１ｉ等の種類によって相違する。なお、センサ２０１ａ〜２０１ｉは、実質的に同一の基本波成分及び高調波成分を含む出力信号２０４を出力するセンサであることが好ましい。例えば、センサ２０１ａ〜２０１ｉは同一の種類であってもよい。However, an error occurs between the calculated temporary angle change amount X _p ⁽⁰⁾ _{and the angle change amount X pideal} of the rotating body 105 that should be obtained from the ideal angle detector 101 (ideally). _Should be X _p = X political). When the rotating body 105 of FIGS. 1A to 2B rotates at a constant speed and is commanded to rotate so that the angle of the rotating body 105 changes, ideally, the rotating body 105 commands as shown in FIG. 4B. Assuming that rotational movement is possible without error with respect to the value, the amount of change in angle X calculated from the output signals 204 from the sensors 201a to 201i as the angle of the rotating body 105 increases due to the rotational movement of the rotating body 105. _p increases linearly, and there is no angle error between the angle of the rotating body 105 and the ideal amount of change in angle X _pideal. In practice, however, the angle error is generated between the angle variation _{X p} and the ideal angle variation _{X Pideal} calculated from the output signal 204 from the sensor 201A～201i. As one of the causes of this angle error, the calculated angle change amount _Xp includes the rotation axis 108 of the rotating body 105 and the central axis of the scale scale 102 in the process of detecting a plurality of scales 103 with the rotation of the rotating body 105. Axis eccentricity with, quality of scale scale 102, aging of angle detector 101, etc. include extra quantities (referred to as mechanical angle error), and the output signal 204 includes sensors 201a-201i. It is due to the inclusion of extra distortion due to the characteristics (angle errors due to the characteristics of the sensors 201a to 201i are referred to as electrical angle errors). More specifically, the cause of the electrical angle error due to the sensors 201a to 201i is that the output signal from the sensors 201a to 201i is a harmonic component in which one scale of the plurality of scales 103 is one cycle primary. And a harmonic component having an order of two or more integral multiples of the fundamental wave component. For example, two pseudo sine wave signals (A phase signal (A ⁽⁰⁾ ) and B phase signal (B ⁽⁰⁾ )) having phases different from each other by 90 ° as shown in FIG. 4A are output signals 204 from the sensors 201a to 201i. When the two pseudo sine wave signals output from the sensors 201a to 201i are output as, the fundamental wave component cos (θ) which is an ideal wave in which one scale of the plurality of scales 103 is one cycle primary. , Sine (θ) and a harmonic component (a component when the order k is an integer of 2 or more) whose order is an integral multiple of 2 or more of the fundamental wave component, as shown below. harmonic components of the sensor 201a~201i is, affects the angle variation X _p when arctangent calculation as described above [Equation 2], as shown in FIG. 4C, the electrical angle error has occurred ..

Here, a _k and b _k are gains of harmonic components of degree k when the amplitude of the first-order fundamental wave component with one scale as one cycle is 1, and φ a _k and _{φ b k} are. This is the phase difference of the harmonic component of order k with respect to the fundamental wave component. _{Incidentally, a k, b k, φa} k, φb k is generally unchanged or even different scales 103, or difference is small as changed. _{Incidentally, a k, b k, φa} k, φb k is characteristic of the sensor 201a~201i and graduated scale 102, dictated by the detection principle, for example, graduated scale 102 if sensor 201a~201i is a light detection type sensor It is determined by the reflection / transmission characteristics of the scale pattern and the sensitivity characteristics of the light receiving part. If the sensors 201a to 201i are semiconductor magnetic resistance type sensors, it is determined by the magnetic resistance characteristics of the semiconductor. When used, the use of spur gears for magnetic detection is determined by the shape characteristics of the teeth. As described above, the harmonic components included in the output signal 204 differ depending on the type of the sensors 201a to 201i and the like. The sensors 201a to 201i are preferably sensors that output an output signal 204 containing substantially the same fundamental wave component and harmonic component. For example, the sensors 201a to 201i may be of the same type.

Further, as described above, the angle error includes not only the electrical angle error due to the characteristics of the sensors 201a to 201i, but also the mounting accuracy of the rotating body 105, the quality of the scale scale 102, and the change with time of the angle detector 101. It also includes mechanical angle errors due to such factors.

Therefore, in order to eliminate the electrical angle error caused by the characteristics of the sensors 201a to 201i, and the mechanical angle error caused by the mounting accuracy of the rotating body 105, the quality of the scale scale 102, the change with time of the angle detector 101, and the like. As described above, the angle detector 101 has a scale scale 102 having a plurality of scales 103 along the circumference of the rotating body 105 with respect to the rotation direction 106, and a sensor along the circumference of the rotating body 105 with respect to the rotation direction 106. At least two sensors 201a to 201i arranged at the planned arrangement locations 202a to 202i are provided, and as shown in FIG. 3, the control unit 203 arithmetically processes the output signals 204 from these sensors 201a to 201i. A signal processing unit 209 is provided.

The control unit 203 sets the input unit 205 for acquiring the output signal 204, the noise filter 206 for removing the noise of the output signal 204, and the output signal 204 prior to the arithmetic processing of the output signal 204 by the signal processing unit 209. An amplifier 207 for amplification and an A / D converter 208 for converting an output signal 204 from an analog value to a digital value may be provided. The output signal 204 converted into a digital value is output to the signal processing unit 209. Further, the control unit 203 may include a storage unit 210 that writes / reads data by the signal processing unit 209. When the output signal 204 is a two-phase signal having different phases, the signal processing unit 209 may be able to adjust the amplitude, offset, and phase difference between the two phases. Even when not adjusted, if the signal characteristics of the sensors are at the same level, these values are considered to be substantially the same, and the electrical angle error can be eliminated by the present invention.

One of the output signals 204 from the sensors 201a to 201i is used as a reference sensor (for example, sensor 201a), and the output signal 204 from the reference sensor and the output signals of the other sensors (for example, sensors 201b to 201i) are used. By obtaining the output signal difference from 204 and obtaining the average value of the obtained output signal difference from other sensors, a calibration value for removing the mechanical angle error can be obtained. By adding and subtracting this calibration value to the calculated temporary angle change amount X _p ^{(0) of} the rotating body 105, the true angle change amount X p can be detected, and the angle detector 101 can detect the _{true angle change amount X p.} , Self-calibrated for mechanical angle error. However, the electrical angle error cannot be eliminated by simply arranging the sensors 201a to 201i along the rotation direction 106 of the rotating body 105. As described above, each output signal 204 of the sensors 201a to 201i has a fundamental wave component in which one scale of the plurality of scales 103 is the first order in one cycle, and an integral multiple of two or more of the fundamental wave components. _{The tentative angular displacement amount X p} ⁽⁰⁾ calculated from the output signal 204, including the harmonic component of order, is 1 of the scale 103 due to one or more harmonic components of the output signal 204. This is because it contains at least one angle error component having an order component that is an integral multiple of the scale with the first order in one cycle.

In order to eliminate the electrical angle error caused by the characteristics of the sensors 201a to 201i, the scale number N of the scale scale 102 along the circumference 107 with respect to the rotation direction 106 of the rotating body 105, and the output signals from the sensors 201a to 201i. Based on the order p of the angle error component in one scale due to the harmonic component included in 204, the number of sensors 201a to 201i arranged along the circumference 107 with respect to the rotation direction 106 of the rotating body 105 is appropriately selected. There is a need. Here, the order of the angle error component is the order of the angle error component, and the order of the angle error component is an integral multiple of the order of one scale of the scale 103, where one cycle is the first order. For example, as shown in FIGS. 1A to 2B, when the scale number of the scale scale 102 is 32 and the order of the angle error component in one scale due to the harmonic component included in the output signal 204 is p. , The number of the sensors 201a to 201i arranged along the circumference 107 is selected based on the scale number 32 of the scale scale 102 and the order p of the angle error component. The order p of the angle error component may be estimated by estimation, or may be determined from the result of calculating the angle error component in one round of the scale scale 102. As a reference, the order p of the angle error component in which the electrical angle error appears remarkably is often 5 or less. Alternatively, the harmonic component may be extracted in advance from the output signals 204 from the sensors 201a to 201i to determine the order p of the angle error component that needs to be removed. Further, when the output signals 204 from the sensors 201a to 201i are input to the signal processing unit 209 via the input unit 205, the noise filter 206, the amplifier 207, the A / D converter 208, etc., as shown in FIG. The number of sensors 201a to 201i may be appropriately selected based on the order p of the angle error component that needs to be removed by extracting the harmonic component included in the signal input to the signal processing unit 209. ..

Further, the provisional angle change amount X _p ⁽⁰⁾ calculated from the output signal 204 is caused by one or more harmonic components included in the output signal 204, and one scale of the scale 103 is set to the first order in one cycle. As a result, a plurality of angle error components may be included, each of which has an order component that is an integral multiple of the order component, and the angle error component may be a plurality of angle error components. The number of sensors 201a to 201i is based on the scale number N of the scale scale 102 and the order of each of the plurality of angular error components, and the sensors 201a to 201a are arranged along the circumference 107 with respect to the rotation direction 106 of the rotating body 105. It is necessary to appropriately select the number of 201i. _{For example, the provisional angular change amount X p} ⁽⁰⁾ calculated from the output signal 204 due to one or more harmonic components of the output signal 204 has two angular errors of order p = 1 and p = 2. In the case where the components are included and as shown in FIGS. 1A to 2B, the rotation direction of the rotating body 105 is based on the scale number 32 of the scale scale 102 and the orders p = 1 and p = 2 of the two angle error components. The number of sensors 201a to 201i arranged along the circumference 107 with respect to 106 is selected.

The number of sensors 201a to 201i may be determined based on an integer that cannot divide the product of the scale number N of the scale scale 102 and the order p of the angle error component in one scale. That is, the specific size of the angle error component and the like appear with individual differences of the scales by the number of scales with respect to one circumference of the rotating body 105, but which order p of the angle error component generated in one scale is. Since the same applies to the scale, the electrical angle error in one round of the scale scale 102 appears as order N _p = N (scale number of scale scale 102) × p (order of the angle error component in one scale), and order N. The number of sensors 201a-201i may be determined based on an integer that is not divisible by _p. As a result, the sensors 201a to 201i can output output signals 204 having different phases. That is, the phases of the fundamental wave component and the harmonic component included in the output signal 204 output by each of the sensors 201a to 201i are different in each of the sensors 201a to 201i.

Specifically, with respect to 256, which is the number of scales N of the scale 103 of the scale scale 102, one scale is used for the temporary angle change amount _Xp ^{(0) due to the harmonic component of the output signal 204.} By multiplying the order p on the assumption that an angle error component having a component of the order p (1 to 10) is included, the order N _p of the electrical angle error in one round of the scale scale 102 is obtained, and the order N is obtained. _{FIG. 5A shows the result of determining p} as x when it is divisible by an integer (5 to 9) as the number of planned sensor placement locations and ◯ when it is not divisible. The number of planned sensor placement locations with a large number of ◯ can eliminate the electrical angle error due to a larger number of components of order p. In the case of FIG. 5A, when the integer is 7 or 9, that is, when the number of planned sensor placement locations is 7 or 9, the number of ○ is 9, and the electrical angle error due to many components of order p is removed. The sensors 201a to 201i are arranged at any of the seven or nine planned sensor arrangement locations 202a to 202i. On the other hand, when the integer is 8, that is, when the number of planned sensor placement locations is 8, the number of ◯ is 0, and the electrical angle error cannot be removed at all. Note that FIG. 5A is an example, and there are no upper or lower limits on the number of scales 103 on the scale scale 102 and the number of planned sensor placement locations.

Figure 1A, as shown in FIG. 2A, the sensor arrangement planned portion 202a~202i are at substantially equal intervals along the circumference 107 with respect to the rotational direction 106 of the rotating body 105, the order of the electrical angle error in the circumference of the graduated scale 102 _{N p} Is set by a number that matches an integer that is not divisible by. As shown in FIG. 5A, when the number of scales N of the scale scale 102 is 256, the number of planned sensor placement locations 202a to 202i is 7 or 9, and 7 or 9 planned sensor placement locations 202a to 202i are rotated. It can be set at substantially equal intervals along the circumference 107 with respect to the rotation direction 106 of the body 105. In FIGS. 1A and 2A, nine sensor placement planned locations 202a to 202i are set along a circumference 107 with respect to the rotation direction 106 of the rotating body 105. The sensors 201a to 201i are arranged at any of the planned sensor placement locations 202a to 202i. In FIG. 1A, the sensors 201a and 201b are arranged at two of the nine planned sensor placement locations 202a to 202i (202a and 202b), and in FIG. 2A, three of the nine sensor placement planned locations 202a to 202i are arranged. Although the sensors 201a, 201b, and 201d are arranged in (202a, 202b, 202d), it does not matter which of the planned sensor arrangement locations 202a to 202i the sensors 202a to 202i are arranged. For example, as shown in FIG. 1A, one sensor 201a and one 201b may be arranged at two adjacent sensor placement planned locations 202a and 202b among the sensor placement planned locations 202a to 202i, and FIG. 1B, FIG. As shown in FIG. 2B, one sensor 201a to 201i may be arranged at each of the planned sensor arrangement locations 202a to 202i.

The electrical angle error appears as order N _p = N (scale number of scale scale 102) × p (order of angle error component in one scale), and since N and p are generally integers, N _p is also an integer. It becomes. However, in reality, due to the fact that the interval between one scale of the scale scale 102 is non-uniform in the plurality of scales 103, etc., the circumference of the rotating body 105 is Fourier transformed as one cycle primary with respect to the angle error. Then, the spectral intensity occurs in a chevron shape in the vicinity of the order of the integer (for example, in the vicinity of the spectral intensity of the order 3), the spectral intensity of the order having a minority point such as 2.9th order, 3.1th order, etc. include). Accordingly, when dividing the order N _p of the electrical angle error in the circumference of the graduated scale 102 in integer as a sensor arrangement planned portion number, the electrical angle error can be removed larger the number of remainder (fraction) in case for the number of order _{N p} increases, based on the remainder, the number of sensors 201a~201i may be determined.

Specifically, as in FIG. 5A, with respect to 256, which is the scale number N of the scale 103 of the scale scale 102, a temporary angle change amount X _p ^{(temporary angle change amount X p () due to the harmonic component of the output signal 204. 0)} is assumed to include an angle error component having a component of order p (1 to 10) on one scale, and is multiplied by a degree p to obtain an order N _{p of an electrical angle error in one circumference of the scale scale 102.} look, the order _{N p,} when divided by an integer (5-9) of the sensor arrangement planned portion number, for example, the remainder is × the case of less than 0.3, the remainder is 0.3 to 0.7 If it is less than, it is marked as ○, if the remainder is 0.7 or more, it is marked as ◎, and it is weighted according to the size of the remainder (for example, × is 0 points, ○ is 1 point, ◎ is 2 points), and the total score is FIG. 5B shows the result of determining the number of planned locations for arranging a large number of sensors. The higher the total score, the more the electrical angle error due to the components of order p can be eliminated. In the case of FIG. 5B, when the integer is 7 or 9, that is, when the number of planned sensor placement locations is 7 or 9, the total score is 9 points, and the electrical angle error due to many components of order p is eliminated. The sensors 201a to 201i are arranged at any of the nine planned sensor placement locations 202a to 202i, as shown in FIGS. 1A to 2B. In this manner, the weighting depending on the magnitude of the remainder when divided by integer of the order N _p of the electrical angle error as a sensor disposed planned portion number in one round of the graduated scale 102, based on the weight of each order The number of sensors 201a to 201i may be determined. Further, when the angle error component in the 1st scale due to the harmonic component of the output signal 204 includes the second-order and fourth-order order components, it is determined as ⊚ in p = 2 and p = 4, so that the sensor. If the number of planned placement locations is selected as 9, secondary and quaternary electrical angle errors can be removed, and as shown in FIGS. 1A to 2B, any of the nine planned placement locations 202a to 202i can be used. , Sensors 201a to 201i are arranged. Thus, based on weighting according to the size of the remainder when dividing the order N _p of the electrical angle error in the circumference of the graduated scale 102 in integer as a sensor arrangement planned portion number, the number of sensors 201a~201i May be determined.

Further, the number of planned sensor placement locations may be determined first, and the scale number N of the scale 103 of the scale scale 102 may be selected. Specifically, the number of planned sensor placement locations is set to 5, and a temporary angle is tentatively caused by the harmonic component of the output signal 204 with respect to 254 to 259, which is the scale number N of the scale 103 of the scale scale 102. The change amount X _p ⁽⁰⁾ is assumed to include an angle error component having a component of order p (1 to 10) on one scale, and is multiplied by the order p to obtain an electrical angle in one circumference of the scale scale 102. When the order N _p of the error is obtained and the order N _p is divided by 5 of the planned number of sensor placement locations, for example, the case where the remainder is less than 0.3 is ×, and the remainder is 0.3 or more and less than 0.7. The case of is ○, the case of the remainder of 0.7 or more is ◎, and the weight is given according to the size of the remainder (for example, × is 0 points, ○ is 1 point, ◎ is 2 points), and the scale scale 102. The result of determining the scale number N in one round is shown in FIG. 5C. The higher the total score, the more electrical angle errors of order p can be eliminated. In the case of FIG. 5C, the electrical angle error of the order p of almost the same number can be removed except when the scale number N in one round of the scale scale 102 is 255, and the five sensor placement planned locations 202a can be removed. Sensors 201a to 201e are arranged in any of ~ 202e. The result of determining the number of scales N in one round of the scale scale 102 has repeatability depending on the number of planned sensor placement locations. For example, the judgment result when the number of planned sensor placement locations is 5 is the same when the number of scales N in one round of the scale scale 102 is 254 ± (5 × integer multiple) (254 in FIG. 5C). In the case of 259, the same determination result is obtained). In this way, it is possible to obtain a determination result having repeatability depending on the number of planned sensor placement locations, regardless of the magnitude of the scale number N in one round of the scale scale 102.

As described above, the scale number N of the scale 103 of the scale scale 102, the order p of the angle error component estimated from the harmonic component included in the output signal 204, or the number of planned sensor placement locations are obtained. , As shown in FIGS. 1A to 2B, the sensors 201a to 201i are arranged at the planned sensor arrangement locations 202a to 202i, and one of the output signals 204 from the sensors 201a to 201i is used as a reference sensor (for example, the sensor 201a). , The output signal difference between the output signal 204 from the reference sensor and the output signal 204 of each other sensor (for example, sensors 201b to 201i) is obtained, and the obtained output signal difference from the other sensors is obtained. By obtaining the average value of, it is possible to obtain a calibration value for removing not only the mechanical angle error but also the electrical angle error. By adding and subtracting this calibration value to the calculated temporary angle change amount X _p ^{(0) of} the rotating body 105, the true angle change amount X p can be detected, and the angle detector 101 can detect the _{true angle change amount X p.} Self-calibrated for angular errors, including mechanical and electrical angular errors. Angle variation X _p after calibration may be as a detection value of the angle detector 101, the motor driving the rotating body 105, the control unit of the rotating body 105, is fed back to an equal, or may be used as a reference angle .. This calibration value may be calculated each time the rotating body 105 rotates and the output signals 204 from the sensors 201a to 201i are input to the signal processing unit 209, or this calibration value may be calculated in advance for one round of the rotating body 105. The calibration value may be calculated by the signal processing unit 209 and stored in the storage unit 210 as a correction table, and this calibration value may be read out from the storage unit 210 when the rotating body 105 rotates.

6A to 6C show the angle errors including the mechanical angle error and the electrical angle error calculated by the signal processing unit 209 based on the output signal 204 from only the sensor 201a of the angle detector 101 of FIG. 1B, respectively. Angle error with respect to the command angle of the rotating body 105 before removal, angle error with respect to the command angle in one scale of the rotating body 105 with a part thereof enlarged, angle error with one circumference of the rotating body 105 as the first order in one cycle The spectral intensity obtained by performing the Fourier transform on the is shown. The angle error due to the component having a long period (small order) with respect to the command angle is due to the mechanical angle error. According to FIG. 6A, in this angle detector 101, the mechanical angle error is The width is about 180 arcsec, with the angle error in which one circumference of the rotating body 105 is the first order in one cycle as the main component. Note that FIG. 6A is a measurement result obtained by extracting only the range of 0 to 180 deg from the measured range of 0 to 360 deg. The angle error due to the component having a short period (large order) with respect to the command angle is due to the electrical angle error, and according to FIG. 6B, the order of the electrical angle error is 1024th order (on the 1st scale). Since the width is about 20 arcsec with the angle error due to the order p = 4) as the main component and the total number of scales of the scale scale 102 is 256, the angle interval 104 of one scale is 1.406 deg, so that it is electrical. The angular error of about 20 arcsec corresponds to about 0.4% of that. According to FIG. 6C, the order component of the angle error is 256th order (1st scale 1st order), 768th order (1st scale 3rd order), 1024th order (1st scale 4th order), and the spectral intensity is large, and the order component of the angle error is It can be seen that the spectral intensity of the 1024th order (1st scale 4th order) is particularly remarkable.

Next, the mechanical angle error and the electrical angle error calculated by the signal processing unit 209 based on the output signals 204 from the sensors 201a to 201i of the angle detector 101 of FIG. 1B are shown in FIGS. 7A to 7C, respectively. After the included angle error is removed, the angle error with respect to the command angle of the rotating body 105, the angle error with respect to the command angle in one scale of the rotating body 105 with a part thereof enlarged, and one cycle of the rotating body 105 are primary. The spectral intensity obtained by performing the Fourier transform on the angle error is shown. Since the sensor arrangement planned portion number (sensor arrangement number) is 9, it is possible to eliminate the angle error, including mechanical angle error and electrical angular error of order N _p can not be divisible by 9, according to FIG. 7A For example, the mechanical angle error of the primary component, which cannot be divided by 9, can be reduced to a width of about 45 arcsec. Note that FIG. 7A is a measurement result obtained by extracting only the range of 0 to 180 deg from the measured range of 0 to 360 deg. According to FIG. 7B, the electrical angle error mainly composed of 1024th order, which cannot be divided by 9, has a width of about 5 arcsec (about 0.1% with respect to 1.406 deg, which is the angle interval 104 of one scale). Can be reduced to. In this way, the angle error including the mechanical angle error and the electrical angle error can be reduced to about 1/4. According to FIG. 7C, the order component of the angle error significantly reduces the spectral intensities of the 256th order (order p = 1st order in the 1st scale), the 768th order (1st scale 3rd order), and the 1024th order (1st scale 4th order). be able to. Further, in FIG. 7A, the 10-ridge component (the 20th-order component with one cycle of the rotating body 105 as the primary) generated during the command angle of 180 deg of the rotating body 105 is due to the mechanical characteristics of the rotating body 105 used. It is known that it is an angle error, and by using the present invention, the mechanical angle error and the electrical angle error are removed from _{the temporary angle change amount Xp} ^{(0) calculated from the output signal 204, and the rotation is performed.} Figure 7A~ Figure 7C shows that can detect the true rotation angle X _p and the angle error of the body 105 with higher accuracy.

Although the above description has been made for a specific embodiment, it is clear to those skilled in the art that the present invention is not limited to this, and various changes and modifications can be made within the scope of the principles of the present invention and the appended claims. is there.

101 Angle detector 102 Scale scale 103 Scale 104 1 Scale angle interval 105 Rotating body 106 Rotation direction 107 One round 108 Rotation axis 201a to 201i Sensor 202a to 202i Sensor placement planned location 203 Control unit 204 Output signal 205 Input unit 206 Noise filter 207 Amplifier 208 A / D converter 209 Signal processing unit 210 Storage unit 211 Display device

Claims (10)

The rotating body includes a rotating body that rotates about the rotation axis, a scale scale having a plurality of scales along the circumference of the rotating body with respect to the rotation direction, and at least two sensors arranged along the circumference. It is an angle detector that detects the amount of angle change due to the rotation of
Each of the at least two sensors outputs a signal according to the amount of change in angle based on the plurality of scales.
The output signal includes a fundamental wave component having one scale of the plurality of scales as the first order in one cycle and a harmonic component having an integral multiple of two or more of the fundamental wave components as the order, and the output. The angular displacement amount calculated from the signal includes at least one angular error component having an order component that is an integral multiple of one scale as one cycle primary due to the harmonic component.
The number of the at least two sensors is determined based on the number of scales on the scale scale and the order component of the at least one angle error component, the angle detector. The at least one angle error component is a plurality of angle error components, and the number of the at least two sensors is determined based on the number of scales on the scale scale and the order component of each of the plurality of angle error components. The angle detector according to claim 1. The number of the at least two sensors is determined based on an integer that cannot divide the product of the number of scales on the scale scale and an integer of 1 or more that is a degree component of the at least one angle error component. Item 2. The angle detector according to item 1 or 2. The angle detector according to claim 3, wherein the number of the at least two sensors is further determined based on the remainder when the product is divided by the indivisible integer. The angle detector according to claim 4, wherein the number of the at least two sensors is further determined based on weighting according to the size of the remainder. A number of planned sensor placement locations corresponding to the indivisible integer are set at approximately equal intervals along the circumference, and at least two sensors are arranged one by one at any of the planned sensor placement locations. The angle detector according to any one of claims 3 to 5. The angle detector according to claim 6, wherein at least two sensors are arranged one by one at two adjacent locations where the sensors are to be arranged. The angle detector according to claim 6, wherein at least one of the two sensors is arranged at each of the planned sensor arrangement locations. The angle detector according to any one of claims 1 to 8, wherein the at least one angle error component differs depending on the type of the at least two sensors. Any one of claims 1 to 9, which is self-calibrated by obtaining the output signal difference between the output signal from one of the at least two sensors and the output signal from the other sensor. The angle detector described in the section.

Images